Neural network predictions of significant coronary artery stenosis in men
Bert A Mobley1, Eliot Schechter, William E Moore
1Department of Physiology, University of Oklahoma Health Sciences Center, College of Medicine, Oklahoma City, OK 73190, USA. bert-mobley@ouhsc.edu
Insights
This study shows artificial neural networks can accurately identify patients with significant coronary stenosis, potentially reducing unnecessary cardiac catheterizations. The model achieved 100% sensitivity in detecting stenosis, aiding clinical decision-making.
Area of Science:
- Cardiology
- Artificial Intelligence
- Medical Diagnostics
Background:
- Coronary stenosis often necessitates invasive cardiac procedures.
- Predictive models can aid in patient selection for interventions.
Purpose of the Study:
- To design and evaluate an artificial neural network (ANN) system for predicting significant coronary stenosis (>50%).
- To assess the potential of ANNs in reducing unnecessary cardiac catheterizations.
Main Methods:
- An ANN was developed using data from 2004 male cardiology patients.
- The network was trained and validated on distinct subsets of the cardiac catheterization database.
- Eleven patient variables were utilized as inputs for the ANN model.
Main Results:
- The ANN achieved 100% sensitivity in identifying patients with significant coronary stenosis in the test set.
- A specificity of 26% was observed for patients without significant stenosis.
- The model demonstrated high accuracy in differentiating between patients who would benefit from intervention.
Conclusions:
- Artificial neural networks show promise as a tool to optimize patient selection for cardiac catheterization.
- The findings suggest ANNs can help reduce the number of non-essential invasive procedures.
Objective:
A neural network system was designed to predict whether coronary arteriography on a given patient would reveal any occurrence of significant coronary stenosis (>50%), a degree of stenosis which often leads to coronary intervention.
Methodology:
A dataset of 2004 records from male cardiology patients was derived from a national cardiac catheterization database. The catheterizations selected for analysis from the database were first-time and elective, and they were precipitated by chest pain. Eleven patient variables were used as inputs in an artificial neural network system. The network was trained on the earliest 902 records in the dataset. The next 902 records formed a cross-validation file, which was used to optimize the training. A third file composed of the next 100 records facilitated the choice of a cutoff number between 0 and 1. The cutoff number was applied to the last 100 records, which comprised a test file.
Results:
When a cutoff of 0.25 was compared to the network outputs of all 100 records in the test file, 12 of 46 (specificity=26%) patients without significant stenosis had outputs
Conclusion:
Artificial neural networks may be helpful in reducing unnecessary cardiac catheterizations.
Related Concept Videos
01:25Imaging Studies for Cardiovascular System VI: Calcium -Scoring CT
01:30Coronary Artery Disease I: Introduction